[SPARK-53907][K8S] Support spark.kubernetes.allocation.maximum
#52615
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
dongjoon-hyun
force-pushed
the
SPARK-53907
branch
from
b4b7b86 to
0a79172
Compare
kazuyukitanimura
approved these changes
Oct 14, 2025
Contributor
kazuyukitanimura
left a comment
kazuyukitanimura
left a comment
There was a problem hiding this comment.
Thank you @dongjoon-hyun LGTM
Member
Author
|
Thank you, @kazuyukitanimura . cc @peter-toth |
Member
|
+1, LGTM |
Member
Author
|
Thank you, @vrozov . |
HyukjinKwon
approved these changes
Oct 15, 2025
Member
Author
|
Thank you, @HyukjinKwon . Merged to master for Apache Spark 4.1.0-preview3. |
dongjoon-hyun
added a commit
that referenced
this pull request
Oct 15, 2025
### What changes were proposed in this pull request? This PR aims to document newly added K8s configurations as a part of Apache Spark 4.1.0 preparation. ### Why are the changes needed? To sync the document with K8s `Config.scala`. For now, three PRs added four configurations. - #51522 - #51811 - #52615 ### Does this PR introduce _any_ user-facing change? No behavior change. This is only adding new configuration documents. ### How was this patch tested? Manual review. ### Was this patch authored or co-authored using generative AI tooling? No. Closes #52618 from dongjoon-hyun/SPARK-53913. Authored-by: Dongjoon Hyun <[email protected]> Signed-off-by: Dongjoon Hyun <[email protected]>
huangxiaopingRD
pushed a commit
to huangxiaopingRD/spark
that referenced
this pull request
Nov 25, 2025
### What changes were proposed in this pull request? This PR aims to support `spark.kubernetes.allocation.maximum`. ### Why are the changes needed? Since we use `AtomicInteger` ID generator, we hit the overflow at `Int.MaxValue`. We had better throw exceptions explicitly in this case because it's highly a malfunctioning situation when a Spark driver tries to create `2147483647 (Int.MaxValue)` executor pods. https://github.com/apache/spark/blob/e05c75e105d5bc9947fb6142f40695ecd5e817e6/resource-managers/kubernetes/core/src/main/scala/org/apache/spark/scheduler/cluster/k8s/ExecutorPodsAllocator.scala#L49 ``` $ jshell | Welcome to JShell -- Version 21.0.8 | For an introduction type: /help intro jshell> var x = new java.util.concurrent.atomic.AtomicInteger(Integer.MAX_VALUE) x ==> 2147483647 jshell> x.incrementAndGet() $2 ==> -2147483648 jshell> x.incrementAndGet() $3 ==> -2147483647 jshell> x.incrementAndGet() $4 ==> -2147483646 jshell> var x = new java.util.concurrent.atomic.AtomicInteger(-1) x ==> -1 jshell> x.incrementAndGet() $6 ==> 0 jshell> x.incrementAndGet() $7 ==> 1 ``` ### Does this PR introduce _any_ user-facing change? Practically no because a normal Spark job and K8s cluster cannot handle `2147483647` executor pod creation. If a user has this case, it means the ID overflow happens and the ID will be rotated to `0` and the executor IDs will be reused. It's already a bug situation. ### How was this patch tested? Pass the CIs with newly added test cases. ### Was this patch authored or co-authored using generative AI tooling? No. Closes apache#52615 from dongjoon-hyun/SPARK-53907. Authored-by: Dongjoon Hyun <[email protected]> Signed-off-by: Dongjoon Hyun <[email protected]>
huangxiaopingRD
pushed a commit
to huangxiaopingRD/spark
that referenced
this pull request
Nov 25, 2025
### What changes were proposed in this pull request? This PR aims to document newly added K8s configurations as a part of Apache Spark 4.1.0 preparation. ### Why are the changes needed? To sync the document with K8s `Config.scala`. For now, three PRs added four configurations. - apache#51522 - apache#51811 - apache#52615 ### Does this PR introduce _any_ user-facing change? No behavior change. This is only adding new configuration documents. ### How was this patch tested? Manual review. ### Was this patch authored or co-authored using generative AI tooling? No. Closes apache#52618 from dongjoon-hyun/SPARK-53913. Authored-by: Dongjoon Hyun <[email protected]> Signed-off-by: Dongjoon Hyun <[email protected]>
shubhluck
pushed a commit
to acceldata-io/spark3
that referenced
this pull request
Dec 11, 2025
### What changes were proposed in this pull request? This PR aims to support `spark.kubernetes.allocation.maximum`. ### Why are the changes needed? Since we use `AtomicInteger` ID generator, we hit the overflow at `Int.MaxValue`. We had better throw exceptions explicitly in this case because it's highly a malfunctioning situation when a Spark driver tries to create `2147483647 (Int.MaxValue)` executor pods. https://github.com/apache/spark/blob/e05c75e105d5bc9947fb6142f40695ecd5e817e6/resource-managers/kubernetes/core/src/main/scala/org/apache/spark/scheduler/cluster/k8s/ExecutorPodsAllocator.scala#L49 ``` $ jshell | Welcome to JShell -- Version 21.0.8 | For an introduction type: /help intro jshell> var x = new java.util.concurrent.atomic.AtomicInteger(Integer.MAX_VALUE) x ==> 2147483647 jshell> x.incrementAndGet() $2 ==> -2147483648 jshell> x.incrementAndGet() $3 ==> -2147483647 jshell> x.incrementAndGet() $4 ==> -2147483646 jshell> var x = new java.util.concurrent.atomic.AtomicInteger(-1) x ==> -1 jshell> x.incrementAndGet() $6 ==> 0 jshell> x.incrementAndGet() $7 ==> 1 ``` ### Does this PR introduce _any_ user-facing change? Practically no because a normal Spark job and K8s cluster cannot handle `2147483647` executor pod creation. If a user has this case, it means the ID overflow happens and the ID will be rotated to `0` and the executor IDs will be reused. It's already a bug situation. ### How was this patch tested? Pass the CIs with newly added test cases. ### Was this patch authored or co-authored using generative AI tooling? No. Closes apache#52615 from dongjoon-hyun/SPARK-53907. Authored-by: Dongjoon Hyun <[email protected]> Signed-off-by: Dongjoon Hyun <[email protected]> (cherry picked from commit 4eacd0b)
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
5 participants
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
What changes were proposed in this pull request?
This PR aims to support
spark.kubernetes.allocation.maximum.Why are the changes needed?
Since we use
AtomicIntegerID generator, we hit the overflow atInt.MaxValue. We had better throw exceptions explicitly in this case because it's highly a malfunctioning situation when a Spark driver tries to create2147483647 (Int.MaxValue)executor pods.spark/resource-managers/kubernetes/core/src/main/scala/org/apache/spark/scheduler/cluster/k8s/ExecutorPodsAllocator.scala
Line 49 in e05c75e
Does this PR introduce any user-facing change?
Practically no because a normal Spark job and K8s cluster cannot handle
2147483647executor pod creation. If a user has this case, it means the ID overflow happens and the ID will be rotated to0and the executor IDs will be reused. It's already a bug situation.How was this patch tested?
Pass the CIs with newly added test cases.
Was this patch authored or co-authored using generative AI tooling?
No.